This week in multidisciplinary engineering and computational research, we highlight ten open-access studies from journals including Transportmetrica B, Results in Engineering, Applied Sciences, and PLoS ONE.
📊 This week at a glance
🌍 African-led research
A new traffic flow model accounts for lane indiscipline and speed differences between vehicle classes, improving predictions for heterogeneous traffic.
Gidey et al. extend second-order multi-class traffic models by incorporating a relative drag function that captures how different vehicles (e.g., cars, trucks) interact when lane discipline breaks down. This changes the assumption that traffic streams are homogeneous and lane-bound, making simulations more realistic for roads where mixed vehicle types share space. For African cities where lane markings are often ignored and vehicle heterogeneity is high, such models can inform better traffic management and infrastructure planning. Published in Transportmetrica B: Transport Dynamics (2025).
Unreliable grid power and back-feeding rules significantly affect the economic viability of rooftop solar-plus-battery systems.
Sanda et al. analyse a grid-connected solar PV system with battery storage at Eagle Square, Abuja, showing that grid outages and the ability to sell excess power back to the grid alter payback periods and returns. This challenges the assumption that grid reliability is a minor factor in solar economics, highlighting that policy on net metering and backup power is crucial. For Nigeria and similar contexts, the findings support tailored incentives for battery storage and clear back-feeding regulations to make solar investments attractive. Published in Results in Engineering (2025).
A Gamma stochastic degradation model estimates railway wheel remaining life using real optical measurements, offering a practical maintenance tool.
Louragli et al. apply a multi-indicator prognostic framework to 80 wheels from ten vehicles, using a CALIPRI C42 profilometer to track wear. This moves beyond fixed maintenance schedules by modelling cumulative, random degradation, potentially reducing unexpected failures and costs. For African rail operators, such data-driven approaches can optimise limited resources and improve safety, though implementation requires access to condition-monitoring technology. Published in Applied Sciences (2025).
The UN Cybercrime Convention may improve cross-border digital evidence sharing, but its impact depends on implementation and existing treaty gaps.
Sarkar et al. assess the UNCC (adopted December 2024) as a response to failures of Mutual Legal Assistance Treaties (MLATs) in delivering timely evidence. This changes the landscape by offering a multilateral framework, but the paper notes that its effectiveness in India and beyond will hinge on domestic adoption and cooperation. For African countries, the UNCC could streamline cybercrime investigations, but requires capacity building and alignment with local laws. Published in Journal of Economic Criminology (2025).
A semi-analytical model compares Marangoni-driven hybrid nanofluid flow with oxytactic microbes over cone, wedge, and plate geometries.
Abbas et al. integrate thermal radiation, viscous dissipation, Dufour-Soret effects, and magnetohydrodynamics using the Xue thermal conductivity model. This provides a unified framework for understanding heat and mass transfer in complex fluids, which could inform cooling systems and biomedical applications. While the study is theoretical, it adds to the knowledge base for African researchers working on nanofluid dynamics and microbial transport. Published in MethodsX (2025).
Metaheuristic tuning of PI controllers for electric vehicle motors improves performance by optimising multiple error metrics simultaneously.
Mashaly et al. use a multi-error ITAE objective function to tune PI controllers in field-oriented control of permanent magnet synchronous motors, enhancing speed and current loop responses. This changes the common practice of optimising a single performance index, leading to more robust EV drive systems. For African EV development, such advanced control strategies can boost efficiency and reliability, though local manufacturing and testing are needed. Published in Machines (2025).
🔬 Global breakthroughs
M2SNet, a multi-scale subtraction network, reduces redundant information in medical image segmentation, improving localisation accuracy.
Zhao et al. propose a network that replaces element-wise addition or concatenation with subtraction to fuse features across scales, addressing blurred edges and mislocalisation in U-shaped models. This could lead to more precise early diagnosis from medical images. For African healthcare, where radiology expertise may be scarce, such automated tools could support screening, but require validation on local datasets. Published in Machine Intelligence Research (2025).
igraph 1.0, a network analysis library, now offers fast, robust performance across Python, R, and Mathematica with a C core.
Antonov et al. release version 1.0 after twenty years of development, enabling reproducible network analysis in multiple programming languages. This lowers barriers for interdisciplinary research, as users can switch languages without losing functionality. For African scholars in social network analysis, epidemiology, or infrastructure planning, igraph provides a free, powerful tool for complex data. Published in PLoS ONE (2025).
Large language model-based intelligent agents are redefining AI by using natural language for reasoning, planning, and tool use.
Cheng et al. review the shift from traditional AI agents to LLM-based agents, which leverage models like GPT for diverse tasks. This changes how agents are designed, moving towards more flexible, language-driven systems. For African researchers, this opens opportunities in natural language processing for local languages and context-aware applications, though computational resources remain a challenge. Published in Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery (2025).
A coupled thermo-hydraulic-mechanical model improves assessment of enhanced geothermal reservoirs by including mechanical deformation.
Li et al. develop a single-fracture model that accounts for how cold fluid injection deforms rock, affecting flow and heat extraction. This addresses the limitation of single-parameter evaluations, offering more accurate predictions of geothermal performance. For African regions with geothermal potential, such models can guide sustainable energy extraction, but require geological data and validation. Published in Energies (2025).
All papers are open access. Explore more Technology & Engineering research on FRELIP · discover open scholarship at frelip.org and search 36,000+ open works at search.frelip.org. FRELIP — born in Nigeria, built for African scholarship, serving the world.